U.S. patent application number 11/552140 was filed with the patent office on 2008-01-10 for fault-isolating sas expander.
This patent application is currently assigned to DOT HILL SYSTEMS CORPORATION. Invention is credited to Ian Robert Davies, George Alexander Kalwitz, James Boyd Lenehan.
Application Number | 20080010530 11/552140 |
Document ID | / |
Family ID | 39226750 |
Filed Date | 2008-01-10 |
United States Patent
Application |
20080010530 |
Kind Code |
A1 |
Davies; Ian Robert ; et
al. |
January 10, 2008 |
FAULT-ISOLATING SAS EXPANDER
Abstract
A SAS expander includes SAS PHYs for transceiving signals with
SAS devices on corresponding SAS links coupled to the SAS PHYs. The
SAS expander includes status registers that provide fault detection
parameters concerning communications on the SAS links. A
microprocessor of the SAS expander identifies faulty communications
on one of the SAS links, based on the fault detection parameters,
and disables a corresponding one of the SAS PHYs coupled to the SAS
link on which the microprocessor identified the faulty
communications. The microprocessor may also report the PHY
disabling to a SAS initiator. The microprocessor may also re-enable
the PHY after corrective action is taken, such as in response to
user input, an indication from a SAS device, or automatically
detecting the corrective action. The expander may also
automatically take the corrective action. The fault detection
parameters may include error counters and corresponding thresholds,
interrupt indicators, and state values.
Inventors: |
Davies; Ian Robert;
(Longmont, CO) ; Kalwitz; George Alexander; (Mead,
CO) ; Lenehan; James Boyd; (Erie, CO) |
Correspondence
Address: |
HUFFMAN LAW GROUP, P.C.
1900 MESA AVE.
COLORADO SPRINGS
CO
80906
US
|
Assignee: |
DOT HILL SYSTEMS
CORPORATION
Carlsbad
CA
|
Family ID: |
39226750 |
Appl. No.: |
11/552140 |
Filed: |
October 23, 2006 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
60804267 |
Jun 8, 2006 |
|
|
|
Current U.S.
Class: |
714/31 ;
714/E11.207 |
Current CPC
Class: |
G06F 11/0793 20130101;
G06F 11/0727 20130101; G06F 11/076 20130101; H04L 1/24
20130101 |
Class at
Publication: |
714/31 |
International
Class: |
G06F 11/00 20060101
G06F011/00 |
Claims
1. A serial-attached-SCSI (SAS) expander for facilitating
communication between SAS devices, comprising: a plurality of SAS
PHYs, for transceiving signals with the SAS devices on a
corresponding plurality of SAS links coupled to said plurality of
SAS PHYs; status registers, coupled to said plurality of SAS PHYs,
configured to provide fault detection parameters concerning
communications on said SAS links; and a microprocessor, coupled to
said status registers, configured to: identify faulty
communications on one of said SAS links based on said fault
detection parameters provided by said status registers; and disable
a corresponding one of said plurality of SAS PHYs coupled to said
one of said SAS links on which said microprocessor identified said
faulty communications.
2. The SAS expander of claim 1, wherein said microprocessor is
configured to forego disabling said one of said plurality of SAS
PHYs if said one of said plurality of SAS PHYs is linked to a
second SAS expander, wherein said second SAS expander is downstream
from a SAS initiator linked to the first SAS expander.
3. The SAS expander of claim 1, wherein said fault detection
parameters provided by said status registers comprise fault counts
associated with said communications on each of said SAS links.
4. The SAS expander of claim 3, wherein said microprocessor is
configured to disable said one of said plurality of SAS PHYs based
on whether one or more of said fault counts exceeds a corresponding
predetermined threshold value.
5. The SAS expander of claim 3, wherein said fault counts comprise
a count of disparity errors received on said SAS link.
6. The SAS expander of claim 3, wherein said fault counts comprise
a count of decode errors detected on said SAS link.
7. The SAS expander of claim 3, wherein said fault counts comprise
a count of dword synchronization losses causing link reset sequence
restarts on said SAS link.
8. The SAS expander of claim 3, wherein said fault counts comprise
a count of dword synchronization failures on said SAS link.
9. The SAS expander of claim 3, wherein said fault counts comprise
a count of CRC dword errors detected on said SAS link.
10. The SAS expander of claim 9, wherein said count of CRC dword
errors detected comprises CRC dword errors detected for IDENTIFY
and OPEN address frames.
11. The SAS expander of claim 9, wherein said count of CRC dword
errors detected comprises CRC dword errors detected while in a SAS
connection.
12. The SAS expander of claim 3, wherein said fault counts comprise
a count of PHY change events generated on said SAS link.
13. The SAS expander of claim 1, wherein said microprocessor is
configured to fetch and execute code instructions to identify said
faulty communications and to disable said corresponding one of said
plurality of SAS PHYs.
14. The SAS expander of claim 1, further comprising: control
registers, coupled to said microprocessor, for controlling said
plurality of SAS PHYs; wherein said microprocessor is configured to
write to said control registers to disable said one of said
plurality of SAS PHYs.
15. The SAS expander of claim 1, further comprising: interrupt
indicators, coupled to said microprocessor, generated by said
plurality of SAS PHYs, for notifying said microprocessor of events
regarding said communications on said SAS links; wherein said
microprocessor is configured to identify said faulty communications
based on said interrupt indicators.
16. The SAS expander of claim 15, further comprising: input
buffers, coupled to said plurality of SAS PHYs, configured to
receive data from said plurality of SAS PHYs; wherein said events
comprise an overflow of said received data out of one of said input
buffers.
17. The SAS expander of claim 15, wherein said events comprise a
transition of one of said plurality of SAS PHYs to a ready
state.
18. The SAS expander of claim 15, wherein said events comprise a
successful execution of a SAS COMINIT sequence with one of said
plurality of SAS PHYs.
19. The SAS expander of claim 1, wherein said fault detection
parameters provided by said status registers comprise a state
associated with said communications on each of said SAS links,
wherein said microprocessor is configured to identify said faulty
communications on said one of said SAS links based on said
state.
20. The SAS expander of claim 19, wherein said state indicates
whether said SAS PHY has been reset a predetermined number of
times.
21. The SAS expander of claim 19, wherein said state indicates
whether a SAS device is connected to said corresponding SAS link
coupled to said SAS PHY.
22. The SAS expander of claim 19, wherein said state indicates
whether a corresponding SAS link is connected to said SAS PHY.
23. The SAS expander of claim 19, wherein said state indicates
whether a corresponding SAS link connected to said SAS PHY was
successfully initialized.
24. The SAS expander of claim 1, further comprising: a memory,
coupled to said microprocessor, for additionally storing said fault
detection parameters, wherein said microprocessor is configured to
identify said faulty communications further based on said fault
detection parameters stored in said memory.
25. A method for increasing data availability in a
serial-attached-SCSI (SAS) system including a SAS expander, having
PHYs, for facilitating communications between a SAS initiator and a
plurality of SAS target storage devices, the method comprising:
identifying, by the SAS expander, faulty communications on a SAS
link connected to a PHY of the SAS expander; and disabling, by the
SAS expander, the SAS expander PHY, in response to said identifying
the faulty communications.
26. The method of claim 25, further comprising: reporting, by the
SAS expander, said disabling.
27. The method of claim 26, wherein said reporting, by the SAS
expander, said disabling comprises reporting, by the SAS expander,
said disabling to the SAS initiator.
28. The method of claim 27, wherein said reporting, by the SAS
expander, said disabling to the SAS initiator comprises reporting,
by the SAS expander, said disabling to the SAS initiator status
information for the SAS initiator to report to a user.
29. The method of claim 26, wherein said reporting, by the SAS
expander, said disabling comprises reporting, by the SAS expander,
said disabling via a SCSI Enclosure Services (SES) page.
30. The method of claim 26, wherein said reporting, by the SAS
expander, said disabling comprises reporting, by the SAS expander,
said disabling via a Serial Management Protocol (SMP) message.
31. The method of claim 25, further comprising: detecting, by the
SAS expander, that action was taken to correct the faulty
communications, after said disabling the SAS expander PHY; and
re-enabling, by the SAS expander, the SAS expander PHY, in response
to said detecting.
32. The method of claim 31, wherein said detecting comprises
receiving information from the SAS initiator that a user took the
corrective action.
33. The method of claim 31, wherein said detecting comprises
receiving information from one of the plurality of SAS target
storage devices that a user took the corrective action.
34. The method of claim 31, wherein the corrective action comprises
replacing a SAS cable coupled to the SAS expander PHY.
35. The method of claim 31, wherein the corrective action comprises
replacing the SAS controller.
36. The method of claim 31, wherein the corrective action comprises
replacing one of the plurality of SAS target storage devices.
37. The method of claim 31, wherein the corrective action comprises
the SAS initiator sending a command to reconfigure a PHY in the SAS
system.
38. The method of claim 31, wherein the corrective action comprises
the SAS expander taking the corrective action.
39. The method of claim 38, wherein the SAS expander taking the
corrective action comprises the SAS expander adjusting analog
settings of the SAS expander PHY.
40. The method of claim 31, wherein said detecting comprises
automatically detecting, by the SAS expander, that a SAS cable
coupled to the SAS expander PHY was replaced.
41. The method of claim 31, wherein said detecting comprises
automatically detecting, by the SAS expander, that the SAS
initiator was replaced.
42. The method of claim 31, wherein said detecting comprises
automatically detecting, by the SAS expander, that said one of the
plurality of SAS target storage devices was replaced.
43. The method of claim 31, wherein said detecting comprises
detecting, by the SAS expander, that the SAS initiator sent a
command to reconfigure a PHY in the SAS system.
44. The method of claim 31, wherein said detecting comprises
detecting, by the SAS expander, the SAS expander taking the
corrective action.
45. The method of claim 44, wherein the SAS expander taking the
corrective action comprises the SAS expander adjusting analog
settings of the SAS expander PHY.
46. The method of claim 25, further comprising: (1) re-enabling, by
the SAS expander, the SAS expander PHY, after said disabling; (2)
determining, by the SAS expander, whether the faulty communications
are persisting, after said re-enabling the SAS expander PHY; (3)
disabling, by the SAS expander, the SAS expander PHY again if the
faulty communications are persisting; and repeating, by the SAS
expander, said steps (1), (2), and (3) until the faulty
communications are no longer persisting at step (2).
47. The method of claim 46, wherein said repeating is performed
periodically.
48. The method of claim 47, wherein said repeating is performed
with a monotonically increasing period.
49. The method of claim 46, further comprising: (4) adjusting, by
the SAS expander, analog settings of the SAS expander PHY, after
said step (3); and repeating, by the SAS expander, said steps (1),
(2), (3), and (4) until the faulty communications are no longer
persisting at step (2).
50. The method of claim 49, further comprising: monitoring, by the
SAS expander, fault detection parameters maintained by the SAS
expander; wherein said identifying is performed by the SAS expander
based on said monitoring.
51. The method of claim 50, wherein said monitoring is performed
periodically by a microprocessor of the SAS expander.
52. The method of claim 51, wherein the SAS initiator comprises a
redundant array of inexpensive disks (RAID) controller.
53. The method of claim 50, wherein the fault detection parameters
comprise error count values related to communications on SAS links
connected to the SAS expander PHYs.
54. The method of claim 53, wherein the fault detection parameters
comprise threshold values for comparing with the error count values
to determine whether the threshold values have been exceeded.
55. The method of claim 50, wherein the fault detection parameters
comprise interrupt indicators to a microprocessor of the SAS
expander PHY that indicate events related to communications on SAS
links connected to the SAS expander PHYs.
56. The method of claim 50, wherein the fault detection parameters
comprise state values related to communications on SAS links
connected to the SAS expander PHYs.
57. The method of claim 49, wherein said repeating is performed no
more than a predetermined number of times.
58. A serial-attached-SCSI (SAS) system, comprising: a SAS
initiator; a plurality of SAS target storage devices; and a SAS
expander, connected to said SAS initiator and to said plurality of
SAS target devices by a plurality of SAS links, said SAS expander
comprising a plurality of PHYs, for coupling to said plurality of
SAS links, wherein said SAS expander is configured to identify
faulty communications on one of said plurality of SAS links and to
responsively disable one of said plurality of PHYs connected to
said identified one of said plurality of SAS links.
59. The system of claim 58, further comprising: a second SAS
initiator; and a second SAS expander, connected to said second SAS
initiator and to said plurality of SAS target devices by a second
plurality of SAS links; wherein said second SAS initiator is
configured to continue to communicate with said plurality of SAS
target devices via said second plurality of SAS links after said
first SAS expander disables said one of said plurality of PHYs
connected to said identified one of said plurality of SAS
links.
60. The system of claim 58, wherein at least two of said plurality
of PHYs are grouped into a wide SAS port for communicating with
said SAS initiator on a wide SAS link comprising a corresponding at
least two of said plurality of SAS links, wherein said SAS expander
is configured to disable one of said at least two of said plurality
of PHYs of said wide SAS port, wherein said SAS expander is
configured to continue to communicate with said SAS initiator via
remaining ones of said at least two of said plurality of PHYs after
said SAS expander disables said one of said at least two of said
plurality of PHYs of said wide SAS port.
Description
CROSS REFERENCE TO RELATED APPLICATION(S)
[0001] This application claims priority to U.S. Provisional
Application No. 60/804,267, filed Jun. 8, 2005, which is hereby
incorporated by reference.
BACKGROUND OF THE INVENTION
Field of the Invention
[0002] The present invention relates in general to the field of
Serial-Attached-SCSI (SAS) systems, and particularly to SAS
expanders.
[0003] Serial-Attached-SCSI (SAS) systems are becoming more and
more common in modern computer systems. SAS systems include SAS
initiator devices and SAS target devices as does its parent, the
Small Computer Systems Interface (SCSI). SAS target devices are
typically storage devices, such as disk drives, that receive
commands from SAS initiator devices, such as SAS host bus adapters
in host computers or SAS I/O controllers in Redundant Arrays of
Inexpensive Disks (RAID) controllers.
[0004] Implementations and uses of SAS are described in detail in
the following documents, each of which is incorporated by reference
in its entirety for all intents and purposes: [0005] "Serial
Attached SCSI--1.1 (SAS-1.1)", Revision 10, Sep. 21, 2005. Working
Draft, Project T10/1601-D, Reference number ISO/IEC 14776-151:200x.
American National Standard Institute.
(http://www.t10.org/ftp/t10/drafts/sas1/sas1r10.pdf) [0006] "Serial
Attached SCSI--2 (SAS-2)", Revision 6, Sep. 22, 2006. Working
Draft, Project T10/1760-D, Reference number ISO/IEC 14776-152:200x.
American National Standard Institute.
(http://www.t10.org/ftp/t10/drafts/sas2/sas2r06.pdf)
[0007] SAS systems are built on point-to-point serial connections
between SAS devices. Each point-to-point connection is referred to
as a link, or lane, and the two endpoints are referred to as a PHY.
A PHY contains a transceiver and electrically interfaces to a link
to communicate with another PHY at the other end of the link. The
link, or lane, includes two differential signal pairs, one in each
direction. A SAS port includes one or more PHYs. A SAS port that
has more than one PHY grouped together is referred to as a wide
port, and the more than one links coupling the two wide ports are
referred to as a wide link. Wide ports and links provide increased
data transfer rates between SAS endpoints.
[0008] The simplest SAS topology is a single SAS initiator having a
SAS port that is connected by a single SAS link to a SAS port of a
single SAS target. However, it is desirable in many applications,
such as a high data availability RAID system, to enable one or more
SAS initiators to communicate with multiple SAS target devices. In
addition to initiators and targets, SAS includes a third type of
device, expanders, which are employed in SAS systems to achieve
more complex topologies. SAS expanders perform switch-like
functions, such as routing, to enable SAS initiators and targets to
communicate via the SAS point-to-point connections.
[0009] The present inventors have observed various problems in
complex topology SAS systems when a component is marginal or goes
bad, such as a SAS device that generates logical errors, improper
PHY analog settings, a bad or marginal PHY, or a bad or marginal
link, which may include bad or marginal cables, connectors, or
printed circuit board assembly traces. Some of the manifestations
of the faulty components include intermittent communication errors
between SAS devices, complete loss of a SAS link, or failure of an
entire SAS domain. Another manifestation is the inability for an
initiator to see a SAS target in the topology due to intermittent
failures that cause a SAS device to work sufficiently well to be
allowed into the topology, but to be sufficiently faulty to prevent
effective communication between SAS devices.
[0010] One method of dealing with these problems is an
initiator-based solution. The initiator may attempt to identify the
faulty component and send a command through the SAS domain to
disable, or bypass, various PHYs in the domain in a trial-and-error
approach until the initiator has isolated the problem. However, the
present inventors have observed some failure scenarios which cannot
be satisfactorily remedied by the initiator-based approach. For
example, assume a component fails in an intermittent fashion, such
as a marginal PHY, that causes a SAS expander to first detect that
a SAS link is operating properly, to subsequently detect that the
link is not operating properly, and to continue this sequence for a
relatively long time. According to the SAS standard, the SAS
expander is required to transmit a BROADCAST primitive on each of
its SAS ports to notify other SAS devices of the change of status
within the SAS domain. Each time a SAS initiator receives the
BROADCAST primitive it is required to perform a SAS discover
process to discover the device type, SAS address, and supported
protocols of each SAS device in the SAS domain and to configure
routing tables within the SAS expanders as needed. The SAS discover
process can take a relatively large amount of time. If the SAS
expander transmits BROADCAST primitives due to the operational to
non-operational link transitions according to a period that is
comparable to the SAS discover process time, then consequently the
SAS initiator may be unable to effectively send commands though the
SAS domain to identify and remedy the problem. Or, even if the
initiator is successful in identifying and fixing the problem, the
SAS domain may have been effectively unavailable for providing user
data transfers for an unacceptable length of time.
[0011] Therefore, what is needed is a solution to improve the data
availability in SAS systems, which are subject to the foregoing
problems.
BRIEF SUMMARY OF INVENTION
[0012] The present invention provides an intelligent SAS expander
that automatically detects faulty communications on one of its
links and isolates the fault by disabling its PHY that is connected
to the link on which it detects the faulty communications. In one
embodiment, the intelligent SAS expander also reports the disabling
of the PHY. In one embodiment, the intelligent SAS expander also
recovers by re-enabling the previously disabled PHY after
corrective action has been taken. The intelligent SAS expander
monitors and analyzes various fault detection parameters to detect
the faulty communications and isolate the faulty component by
disabling the PHY.
[0013] In one aspect, the present invention provides a
serial-attached-SCSI (SAS) expander for facilitating communication
between SAS devices. The SAS expander includes a plurality of SAS
PHYs, for transceiving signals with the SAS devices on a
corresponding plurality of SAS links coupled to the plurality of
SAS PHYs. The SAS expander also includes status registers, coupled
to the plurality of SAS PHYs, configured to provide fault detection
parameters concerning communications on the SAS links. The SAS
expander also includes a microprocessor, coupled to the status
registers. The microprocessor identifies faulty communications on
one of the SAS links based on the fault detection parameters
provided by the status registers. The microprocessor also disables
a corresponding one of the plurality of SAS PHYs coupled to the one
of the SAS links on which the microprocessor identified the faulty
communications.
[0014] In another aspect, the present invention provides a method
for increasing data availability in a serial-attached-SCSI (SAS)
system including a SAS expander, having PHYs, for facilitating
communications between a SAS initiator and a plurality of SAS
target storage devices. The method includes the SAS expander
identifying faulty communications on a SAS link connected to a PHY
of the SAS expander. The method also includes the SAS expander
disabling the SAS expander PHY, in response to identifying the
faulty communications.
[0015] In another aspect, the present invention provides a
serial-attached-SCSI (SAS) system. The system includes a SAS
initiator, a plurality of SAS target storage devices, and a SAS
expander, connected to the SAS initiator and to the plurality of
SAS target devices by a plurality of SAS links. The SAS expander
includes a plurality of PHYs, for coupling to the plurality of SAS
links. The SAS expander identifies faulty communications on one of
the plurality of SAS links and responsively disables one of the
plurality of PHYs connected to the identified one of the plurality
of SAS links.
[0016] An advantage of the present invention is that it provides a
more direct, deterministic, and robust method for responding to
faults in SAS systems than a SAS initiator-based solution. In
particular, the intelligent SAS expanders described herein may
directly disable a PHY associated with a detected fault by directly
accessing registers within the SAS expander that control the PHY,
without relying on the transmission of commands through the SAS
domain to disable the PHY as required by an initiator-based
solution, which, as described herein, might be ineffective or
unacceptable with some failure modes.
BRIEF DESCRIPTION OF THE DRAWINGS
[0017] FIG. 1 is a block diagram of a SAS system according to the
present invention.
[0018] FIG. 2 is a block diagram illustrating in more detail the
SAS expander of FIG. 1.
[0019] FIG. 3 is a block diagram illustrating fault detection
parameters according to the present invention.
[0020] FIG. 4 is a flowchart illustrating operation of the SAS
system of FIG. 1 according to one embodiment of the present
invention.
[0021] FIG. 5 is a flowchart illustrating operation of the SAS
system of FIG. 1 according to an alternate embodiment of the
present invention.
[0022] FIG. 6 is a flowchart illustrating operation of the SAS
system of FIG. 1 according to an alternate embodiment of the
present invention.
[0023] FIG. 7 is a flowchart illustrating operation of the SAS
system of FIG. 1 according to an alternate embodiment of the
present invention.
[0024] FIG. 8 is a flowchart illustrating operation of the SAS
system of FIG. 1 according to an alternate embodiment of the
present invention.
DETAILED DESCRIPTION
[0025] Referring now to FIG. 1, a block diagram of a SAS system 100
according to the present invention is shown. The SAS system 100
includes two host computers 108 each coupled to two SAS-based RAID
controllers 104, via a host interconnect such as Ethernet,
FibreChannel, or the like. Each RAID controller 104 is coupled to a
corresponding SAS expander 102 via a wide SAS link 112. In the
embodiment of FIG. 1, the wide SAS links 112 between the RAID
controllers 104 and the SAS expanders 102 are 4.times. wide. Each
of the SAS expanders 102 is coupled to a plurality of SAS disks 106
via a corresponding SAS link 112. In the embodiment of FIG. 1, the
SAS links 112 between the SAS expanders 102 and the SAS disks 106
are narrow (i.e., 1.times.) SAS links 112. The RAID controllers
104, SAS expanders 102, and SAS disks 106 are enclosed in an
enclosure 114. Additionally, the SAS expanders 102 within the
enclosure 114 are connected via a SAS link 112. In one embodiment,
the inter-expander SAS link 112 is a 4.times. wide link. The
inter-expander SAS link 112 advantageously provides a second SAS
pathway for each RAID controller 104 to communicate with each of
the SAS disks 106 through the SAS expander 102 that is directly
connected to the other RAID controller 104.
[0026] Advantageously, the SAS expanders 102 of the SAS system 100
are intelligent SAS expanders 102 that include the ability to
identify faulty communications on a SAS link 112 connected to one
of the SAS expander 102 PHYs. Furthermore, the intelligent SAS
expanders 102 include the ability to disable the identified PHY to
isolate the faulty component, which may be the PHY itself, from the
rest of the SAS system 100. Additionally, the intelligent SAS
expanders 102 include the ability to report the disabled PHY.
Finally, the intelligent SAS expanders 102 include the ability to
recover from faulty condition. In one embodiment, a user notifies
the SAS expander 102 that corrective action has been taken, such as
replacing the faulty component (e.g., faulty cable, faulty SAS disk
106, or other faulty component), and the SAS expander 102
responsively repairs the communication between the SAS expander 102
and the other device by re-enabling the previously disabled PHY. In
one embodiment, the SAS expander 102 is intelligent enough to
automatically detect that a user has remedied the fault, and
responsively re-enables the PHY. In one embodiment, the SAS
expander 102 is intelligent enough to automatically take action to
remedy the fault, such as by adjusting the PHY analog settings
until reliable communications are restored. The SAS expander 102
includes a microprocessor that monitors status registers associated
with the PHYs to identify faulty communications on a SAS link 112,
writes to control registers to disable and re-enable the PHYs, and
performs the reporting function, as discussed in detail below.
[0027] The SAS system 100 of FIG. 1 also includes second and third
enclosures 114 similar to the first enclosure 114 described above;
however, the second and third enclosures 114 do not include the
RAID controllers 104, and are employed for enclosing only SAS disks
106 and two SAS expanders 102. Each SAS expander 102 in the first
enclosure 114 is linked to a corresponding one of the SAS expanders
102 in the second enclosure 114 via a corresponding wide SAS link
112. Similarly, each SAS expander 102 in the second enclosure 114
is linked to a corresponding one of the SAS expanders 102 in the
third enclosure 114 via a corresponding wide SAS link 112. In one
embodiment, each enclosure 114 may enclose up to 12 SAS disks 106,
in addition to the RAID controllers 104, SAS expanders 102, power
supplies, cooling systems, management controllers, and other
components as are well known in the storage system industry.
[0028] Advantageously, the SAS system 100 of FIG. 1 is arranged in
a redundant manner to increase fault-tolerance of the SAS system
100. In particular, each SAS disk 106 is accessible by each of the
RAID controllers 104 so that if a RAID controller 104, SAS link
112, or SAS expander 102 fails, the hosts 108 may continue to
access the SAS disks 106 via a surviving RAID controller 104 or SAS
pathway. The SAS system 100 of FIG. 1 is intended as an example of
a SAS system 100 in which the present invention may be employed.
However, the present invention is not limited to the system
configuration shown in FIG. 1. Rather, the present invention may be
employed in various SAS topologies that include SAS expanders. In
some embodiments, the hosts 108 may also be included in the first
enclosure 114.
[0029] As mentioned above, the SAS links 112 may include various
components, such as cables, connectors, and printed circuit board
assemblies that include signal conductors. In one embodiment, the
SAS expander 102 comprises a PM8388 SXP 24.times.3G 24-port SAS
expander available from PMC-Sierra, Inc., of Santa Clara, Calif.,
in which the present inventors have modified the code 222
(discussed below) to perform the fault identification, isolation,
reporting, and repairing steps of the present invention described
herein. In other embodiments, the SAS expander 102 comprises a
modified version of the following PMC-Sierra models: PM8387 SXP
36.times.3G 36-port SAS expander, PM8399 SXP 24.times.3GSEC 24-port
SAS expander, or PM8398 SXP 36.times.3GSEC 24-port SAS
expander.
[0030] Although FIG. 1 illustrates a SAS system 100 including SAS
disks 106, the present invention may be employed in a SAS system
100 including SATA disks 106, which are interoperable with SAS
disks within a SAS domain. In particular, the SATA Tunneled
Protocol (STP) provides a means for SAS/SATA initiators to
communicate with SATA disks over the SAS hardware
infrastructure.
[0031] Referring now to FIG. 2, a block diagram illustrating in
more detail the SAS expander 102 of FIG. 1 is shown. The SAS
expander 102 includes a microprocessor 202 coupled to a memory 204
and a plurality of SAS ports 216. The memory 204 stores code 222
instructions that are fetched and executed by the microprocessor
202 to accomplish the fault identification, isolation, reporting,
and repairing steps described herein. The memory 204 also stores
threshold values 224 which the microprocessor 202 compares with
counter values 392 (described below with respect to FIG. 3) to
detect faulty communications, as discussed below.
[0032] Each SAS port 216 includes one or more SAS PHYs 208 that is
connected to one of the SAS links 112 of FIG. 1. As shown, some of
the SAS ports 216 are wide SAS ports 216 and some are narrow SAS
ports 216. Each SAS port 216 also includes a SAS SERDES circuit
(not shown).
[0033] The SAS expander 102 also includes a set of control and
status registers (CSRs) 206 associated with each PHY 208, which the
microprocessor 202 reads and writes to monitor fault detection
parameters 300 (described below with respect to FIG. 3) and to
control the PHYs 208, such as to disable and enable the PHYs 208.
The fault detection parameters 300 and their use are described in
more detail below with respect to the remaining Figures. In
addition, the fault detection parameters 300 include interrupt
indicators 218 from the PHYs 208 that are provided to notify the
microprocessor 202 of events related to communication on the SAS
links 112.
[0034] The SAS expander 102 also includes multiplexed data paths
(such as a crossbar) and switching circuitry (not shown) that
interconnect the various PHYs 208 to enable them to transfer
commands and data from one PHY 208 to another to perform the
switching function of the SAS expander 102. The SAS expander 102
may also include buffering circuits associated with each of the
PHYs 208 for buffering the commands and data when received in a
port 216 and when waiting to be transmitted out a port 216. The
commands and data are routed through the network between the ports
216 based on routing table information, which in one embodiment is
stored in the memory 204.
[0035] Referring now to FIG. 3, a block diagram illustrating fault
detection parameters 300 according to the present invention is
shown. The fault detection parameters 300 include the interrupt
indicators 218, values stored in the CSRs 206, and values stored in
the memory 204 of FIG. 2. The fault detection parameters 300 may be
categorized generally as counters 392, interrupt indicators 218,
states 394, thresholds 396 stored in the CSRs 206 and the
thresholds 224 stored in the memory 204 of FIG. 2. The fault
detection parameters 300 also include a performance monitoring
period 372 stored in a CSR 206, whose use is described below.
[0036] The microprocessor 202 maintains a corresponding threshold
396 for each of the counters 392. Some of the thresholds 396 are
stored in the CSRs 206, namely the disparity error interval
threshold 362 and the code violation error interval threshold 364,
and the SAS expander 102 hardware automatically compares them with
the corresponding counter 392 value and generates an interrupt if
the threshold is exceeded. The thresholds 224 corresponding to the
other counter 392 values are stored in the memory 204, and the
microprocessor 202 periodically compares the counter 392 values, or
accumulated counts derived from the periodically sampled counter
392 values, with the thresholds 224 to identify faulty
communications on the SAS links 112.
[0037] The counters 392 include an invalid DWORD count 302, which
indicates the number of invalid DWORDs received outside PHY reset
sequences; a disparity error count 304, which indicates the number
of running disparity errors received outside PHY reset sequences; a
code violation count 306, which indicates the number of times a
decode error was detected on a bit stream; a loss of DWORD
synchronization count 308, which indicates the number of times the
PHY 208 has restarted the link reset sequence because it lost dword
synchronization (i.e., the number of times the PHY 208 went from
PHY ready state to COMINIT state); a PHY reset failed count 312,
which indicates the number of times the PHY 208 has failed to
obtain dword synchronization during final SAS speed negotiation; a
CRC error count 314, which indicates the number of CRC DWORD errors
detected for received IDENTIFY and OPEN address frames; an in
connection CRC error count 316, which indicates the number of in
connection CRC errors; and a PHY change count 318, which indicates
the number of PHY change events that have been generated.
[0038] The interrupt indicators 218 include a PHY ready interrupt
322, which indicates the PHY 208 has finished initialization and is
ready to transmit and receive data (A PHY 208 becomes ready only
after COMINIT has been detected); a COMINIT interrupt 324, which
indicates a valid COMINIT out of band (OOB) sequence has been
successfully negotiated; a elastic store overflow interrupt 326,
which indicates a valid DWORD was received and the internal elastic
store, or buffer, is full; a disparity error interrupt 328, which
indicates the disparity error interval threshold 362 has been
exceeded during the number of clock cycles specified in the
performance monitoring period 372; a code violation error interrupt
332, which indicates the code violation error interval threshold
364 has been exceeded during the number of clock cycles specified
in the performance monitoring period 372; a DWORD synchronization
loss interrupt 334, which indicates DWORD synchronization on the
PHY 208 was lost and consequently the PHY 208 has restarted the
link reset sequence.
[0039] The states 394 include a link connected state 342, which
indicates whether the port 216 is in a connected state; a DWORD
synchronization lost state 344, which indicates the PHY 208 has
currently lost DWORD synchronization; an init passed state 346,
which indicates whether the port 216 has successfully completed the
link initialization sequence; a device present state 348, which
indicates whether a device is connected to the PHY 208; an attached
device type state 352, which indicates whether a SAS or SATA device
was detected as being connected; a rate state 354, which indicates
whether the final negotiated line rate is 1.5 or 3.0 Gbits/sec; a
PHY reset limit saturation state 356, which indicates that the PHY
208 reset threshold has been reached.
[0040] In one embodiment, the SAS expander 102 is configured to
receive from the RAID controllers 104 SCSI Enclosure Services (SES)
pages that set and get the various fault detection parameters 300,
that get the status of the PHYs 208, and that directly enable or
disable individual PHYs 208. In one embodiment, control and status
information, such as SES pages, may be sent via an out-of-band
communication path between the SAS expanders 102 within an
enclosure 114, such as an 12C connection or other communication
path. The out-of-band communication path may be advantageously
employed if the SAS expander 102 has disabled all PHYs 208
connecting the SAS expander 102 to an upstream SAS expander 102,
such as might occur if the SAS cable connecting them is faulty. The
disabling SAS expander 102 may communicate to the other SAS
expander 102 in the enclosure 114 status information indicating
that it has disabled the PHYs 208. In this situation, to avoid
rebooting, the user may cause the other SAS expander 102 in the
enclosure 114 to broadcast an SES page via the out-of-band
communication path to the PHY-disabled SAS expander 102 instructing
the SAS expander 102 to re-enable the disabled PHYs 208 after the
cable has been replaced. The out-of-band communication path is
particularly useful for the SAS expanders 102 within an enclosure
114 that do not have an inter-expander SAS link 112, which may not
be present because the SAS specification does not allow loops
within the SAS topology. Furthermore, the SAS expander 102 includes
default values of the fault detection parameters 300 that are
stored in a non-volatile memory of the SAS expander 102 and that
are employed at boot time of the SAS expander 102. The default
values may be modified by the RAID controllers 104 or by the
microprocessor 202 during operation.
[0041] Referring now to FIG. 4, a flowchart illustrating operation
of the SAS system 100 of FIG. 1 according to one embodiment of the
present invention is shown. Flow begins at block 402.
[0042] At block 402, the microprocessor 202 of FIG. 2 of the SAS
expander 102 of FIG. 1 periodically monitors the fault detection
parameters 300 of FIG. 3 of each of its PHYs 208. The interrupt
indicators 218 are polled by the microprocessor 202. In one
embodiment, the interrupts indicators 218 are received
asynchronously by the microprocessor 202 as interrupt request
signals. In one embodiment, in response to monitoring the fault
detection parameters 300, the microprocessor 202 also updates a
database that it maintains for providing status information to the
RAID controllers 104. In one embodiment, the status information is
provided via SES pages. In one embodiment, in response to
monitoring the fault detection parameters 300, the microprocessor
202 also maintains and updates accumulated error counts stored in
the memory 204 over multiple monitoring periods. In one embodiment,
the monitoring period is determined by a timer interrupt to the
microprocessor 202. Flow proceeds to block 404.
[0043] At block 404, the microprocessor 202 identifies faulty
communications on a SAS link 112 connected to one of its PHYs 208
based on the monitoring at block 402. The microprocessor 202
analyzes the fault detection parameters 300 according to isolation
rules embodied in the code 222 for fault indications to determine
whether there is a need to disable a PHY 208. The identification of
the faulty communications may include various criteria as discussed
herein. An isolation rule may be triggered by one or more of the
various counts exceeding a threshold, by detection that a PHY 208
has reached one or more particular states, that one or more
particular events have occurred as indicated by one or more of the
interrupt indicators 218, and various combinations thereof. In one
embodiment, the microprocessor 202 only identifies faulty
communications related to a PHY 208 if the PHY 208 is enabled. In
one embodiment, the microprocessor 202 only identifies faulty
communications related to a PHY 208 if isolation is allowed for the
PHY 208. In one embodiment, the SAS expander 102 receives SES pages
from the RAID controllers 104 to selectively enable and disable
individual PHYs 208 and to selectively allow and disallow isolation
of individual PHYs 208. Flow proceeds to block 406.
[0044] At block 406, the microprocessor 202 writes to a control
register 206 to disable the PHY 208 identified at block 404. Flow
proceeds to block 408.
[0045] At block 408, the SAS expander 102 reports the fact that the
PHY 208 was disabled to one or both of the RAID controllers 104. In
one embodiment, the SAS expander 102 also reports the reason the
PHY 208 was disabled. In one embodiment, the SAS expander 102 also
reports all threshold values used by the SAS expander 102 to make a
determination to disable the PHY 208. In one embodiment, the SAS
expander 102 reports by transmitting an SES diagnostic page to the
RAID controller 104. In one embodiment, the SAS expander 102
reports by transmitting a Serial Management Protocol (SMP) message
to the RAID controller 104. In one embodiment, the SAS expander 102
provides an interface to the RAID controllers 104 to enable the
RAID controllers 104 to obtain the status of each PHY 208 and the
current error counts, state, and events described herein. Flow
proceeds to block 412.
[0046] At block 412, the RAID controller 104 reports that the PHY
208 was disabled to a user. In one embodiment, the RAID controller
104 reports to the user via a management interface. In one
embodiment, the RAID controller 104 reports to the user by
reporting to one or both of the hosts 108, which in turn notify the
user. Flow proceeds to block 414.
[0047] At block 414, assuming the disabled PHY 208 is part of a
wide port 216, communications between the SAS expander 102 port and
the SAS device connected to the port 216 continue via the remaining
PHYs 208 of the port 216 and associated SAS links 112 that are
functioning properly. It is noted that the SAS system 100 may
experience a proportionally lower data throughput due to the
disabled PHY 208 and its respective SAS link 112. However,
advantageously, by disabling the PHY 208 associated with the faulty
SAS link 112 (or the PHY 208 itself may have been faulty), the
likelihood that the SAS system 100 will continue functioning
normally is increased, thereby improving the availability of the
data on the SAS disks 106 to the hosts 108, rather than
experiencing the various problems discussed herein. Flow ends at
block 414.
[0048] Referring now to FIG. 5, a flowchart illustrating operation
of the SAS system 100 of FIG. 1 according to an alternate
embodiment of the present invention is shown. The operation of the
SAS system 100 as described in FIG. 5 assumes that the disabled PHY
208 was part of a narrow port 216 (rather than a wide port 216 as
assumed with respect to FIG. 4) such that communication between the
SAS expander 102 and the SAS device that was connected to the PHY
208 that was disabled at block 406 is no longer possible via the
narrow SAS link 112 between the SAS expander 102 and the SAS
device. Additionally, FIG. 5 does not describe attempts to recover
normal functioning of the faulty SAS link 112, such as is described
with respect to FIGS. 6 through 8. In situations where recovery of
normal functioning is performed, the SAS system 100 may continue to
operate as described in FIG. 4 or 5 (depending upon whether the
disabled PHY 208 was part of a wide or narrow port 216) until the
recovery of normal functioning is achieved.
[0049] Flow begins at block 402. Blocks 402 through 412 of FIG. 5
are the same as like-numbered blocks of FIG. 4 and for the sake of
brevity are not described again here. Flow proceeds from block 412
of FIG. 5 to block 514.
[0050] At block 514, the hosts 108 continue to access the SAS disks
106 implicated by the PHY 208 disabled at block 406 via an
alternate pathway that does not include the disabled PHY 208. With
respect to the SAS system 100 of FIG. 1, the hosts 108 will
communicate with the SAS disks 106 via the other RAID controller
104. Flow ends at block 512.
[0051] Referring now to FIG. 6, a flowchart illustrating operation
of the SAS system 100 of FIG. 1 according to an alternate
embodiment of the present invention is shown. FIG. 6 describes
operation of the SAS system 100 of FIG. 1 in which the SAS expander
102 additionally re-enables the previously disabled PHY 208 in
response to user input that the fault has been corrected.
[0052] Flow begins at block 402. Blocks 402 through 412 of FIG. 6
are the same as like-numbered blocks of FIG. 4 and for the sake of
brevity are not described again here. Flow proceeds from block 412
of FIG. 6 to block 614.
[0053] At block 614, the user takes action to correct the faulty
component in response to the reporting of the disabled PHY 208 at
block 412. Examples of action that the user may take to correct the
faulty component include, but are not limited to, replacing a
cable, replacing a connector, replacing a SAS disk 106, replacing a
SAS expander 102, replacing a RAID controller 104, and
reconfiguring a PHY 208, such as to adjust its analog settings.
Flow proceeds to block 616.
[0054] At block 616, the user notifies one of the RAID controllers
104 that he has taken the corrective action at block 614. In one
embodiment, the user notifies the RAID controller 104 via a
management interface. In one embodiment, the user notifies one of
the hosts 108, which in turn notifies the RAID controller 104. Flow
proceeds to block 618.
[0055] At block 618, the RAID controller 104 notifies the SAS
expander 102 that the corrective action was taken. In one
embodiment, the SAS expander 102 is notified by receiving a SCSI
Enclosure Services (SES) diagnostic page from the RAID controller
104. In one embodiment, the SAS expander 102 is notified by
receiving a Serial Management Protocol (SMP) message from the RAID
controller 104. In one embodiment, the RAID controller 104 notifies
the SAS expander 102 by explicitly instructing the SAS expander 102
to re-enable the PHY 208. Flow proceeds to block 622.
[0056] At block 622, the microprocessor 202 writes to a control
register 206 to re-enable the PHY 208 that was previously disabled
at block 406, in response to the notification that the corrective
action was taken. Flow ends at block 622.
[0057] In one embodiment, the microprocessor 202 foregoes disabling
the PHY 208 at block 406 if the PHY 208 is linked to another SAS
expander 102 that is downstream from a RAID controller 104 linked
to the SAS expander 102 that detected the fault. This
advantageously simplifies recovery of certain failure modes on a
SAS topology involving cascaded SAS expanders 102, such as the SAS
system 100 of FIG. 1. For example, assume a cable is faulty that
connects a SAS expander 102 in each of two of the enclosures 114 of
FIG. 1 and both of the SAS expanders 102 detect the fault and
disable their respective PHYs 208. In this example, to recover
operation of the SAS link 112 once the cable is replaced may
require coordination between the two SAS expanders 102 and
potentially between the two RAID controllers 104. In contrast, by
foregoing disabling the PHY 208 at block 406 if the PHY 208 is
linked to a downstream SAS expander 102, the requirement for
coordination is avoided.
[0058] Referring now to FIG. 7, a flowchart illustrating operation
of the SAS system 100 of FIG. 1 according to an alternate
embodiment of the present invention is shown. FIG. 7 describes
operation of the SAS system 100 of FIG. 1 in which the SAS expander
102 additionally re-enables the previously disabled PHY 208 in
response to automatically detecting that the fault has been
corrected.
[0059] Flow begins at block 402. Blocks 402 through 412 and 614 of
FIG. 7 are the same as like-numbered blocks of FIG. 6 and for the
sake of brevity are not described again here. Flow proceeds from
block 614 of FIG. 7 to block 716.
[0060] At block 716, the microprocessor 202 automatically detects
that the user took the corrective action at block 614. In one
embodiment, the user corrective action automatically detected by
the microprocessor 202 is a user replacing a cable. The
microprocessor 202 automatically detects the cable replacement by
detecting a change of state from link not connected to link
connected via the link connected state 342 fault detection
parameter 300. In one embodiment, the user corrective action
automatically detected by the microprocessor 202 is a user
replacing a SAS disk 106 or a SATA disk. The microprocessor 202
automatically detects the disk replacement by detecting a change of
state from device not present to device present via the device
present state 348 fault detection parameter 300 and detects whether
the replaced disk is a SAS disk or a SATA disk via the attached
device type state 352. In one embodiment, the user corrective
action automatically detected by the microprocessor 202 is a user
replacing a SAS expander 102. The microprocessor 202 automatically
detects the SAS expander 102 replacement by detecting a change of
state from device not present to device present via the device
present state 348 fault detection parameter 300 of a PHY 208
connected to the replaced SAS expander 102 via the inter-expander
SAS link 112 of FIG. 1 or via receiving status on the out-of-band
communication path discussed above with respect to FIG. 3. In one
embodiment, the user corrective action automatically detected by
the microprocessor 202 is a user replacing a RAID controller 104.
The microprocessor 202 automatically detects the RAID controller
104 replacement by detecting a change of state from device not
present to device present via the device present state 348 fault
detection parameter 300 of a PHY 208 connecting the SAS expander
102 to the replaced RAID controller 104. The above are provided as
examples of automatically detected user corrective action performed
by the microprocessor 202; however, the present invention is not
limited to the embodiments described above, but may be employed for
other user corrective action. Flow proceeds to block 718.
[0061] At block 718, the microprocessor 202 writes to a control
register 206 to re-enable the PHY 208 that was previously disabled
at block 406, in response to the automatic detection at block 716
that the corrective action was taken by the user. Flow ends at
block 718.
[0062] Referring now to FIG. 8, a flowchart illustrating operation
of the SAS system 100 of FIG. 1 according to an alternate
embodiment of the present invention is shown. FIG. 8 describes
operation of the SAS system 100 of FIG. 1 in which the SAS expander
102 automatically takes action to attempt to correct the faulty
condition and recover communications on the previously disabled PHY
208.
[0063] Flow begins at block 402. Blocks 402 through 412 of FIG. 8
are the same as like-numbered blocks of FIG. 4 and for the sake of
brevity are not described again here. Flow proceeds from block 412
of FIG. 8 to block 814.
[0064] At block 814, the microprocessor 202 automatically takes
corrective action. In one embodiment, the automatic corrective
action taken by the microprocessor 202 is to automatically adjust
the PHY 208 analog settings, which may cause the SAS link 112 to
start functioning properly if, for example, the cable length has
been changed since the last time the PHY 208 analog settings were
set. Flow proceeds to block 816.
[0065] At block 816, the microprocessor 202 writes to a control
register 206 to re-enable the PHY 208 that was previously disabled
at block 406. Flow proceeds to decision block 818.
[0066] At decision block 818, the microprocessor 202 determines
whether normal communications have been restored on the SAS link
112 after re-enabling the PHY 208. If so, flow ends; otherwise,
flow proceeds to block 822.
[0067] At block 822, the microprocessor 202 disables the PHY 208
again. Flow returns from block 822 to block 814.
[0068] In one embodiment, the microprocessor 202 maintains a retry
count threshold, and once the microprocessor 202 has performed the
steps in the loop at blocks 814 to 822 a number of times that
exceeds the retry threshold, the microprocessor 202 leaves the PHY
208 disabled and stops trying to automatically repair the fault
until it detects an event indicating that it should re-enable the
PHY 208.
[0069] In one embodiment, the microprocessor 202 increases the
period of the steps performed in the loop at blocks 814 to 822 each
time it disables the PHY 208 at block 822 in order to reduce the
number of SAS discover processes that must be performed in response
to the PHY 208 disabling/re-enabling. A management application
client performs a SAS discover process to discover all the SAS
devices and expander devices in the SAS domain (i.e., determining
their device types, SAS addresses, and supported protocols). A SAS
initiator device uses this information to determine SAS addresses
to which it is able to establish connections. A self-configuring
expander device uses this information to fill in its expander route
table. Additionally, if there are multiple disabled PHYs 208 that
need re-enabling, then the microprocessor 202 re-enables all of the
disabled PHYs 208 at the same time in order to further reduce the
number of SAS domain discover processes that must be performed.
[0070] Although the present invention and its objects, features,
and advantages have been described in detail, other embodiments are
encompassed by the invention. For example, although embodiments
have been described in which the SAS initiators are RAID
controllers, other embodiments are contemplated in which other
types of SAS initiators are employed. Furthermore, although
embodiments have been described in which the SAS targets are disk
drives, other embodiments are contemplated in which other types of
SAS storage devices are employed.
[0071] Finally, those skilled in the art should appreciate that
they can readily use the disclosed conception and specific
embodiments as a basis for designing or modifying other structures
for carrying out the same purposes of the present invention without
departing from the spirit and scope of the invention as defined by
the appended claims.
* * * * *
References